In today's world, multimedia is prolific and users can experience multimedia in a number of ways. Of particular popularity among users, especially those seeking professional singing careers, is the ability to create an audio or a combination video and audio recording of oneself performing (e.g., singing or playing) a composition previously-recorded by a professional artist. Users then post the performance online for other users, e.g., a coach or professional contact, to view. Systems facilitating posting of such content typically match the composition performed by the user with a composition previously recorded by the professional artist. Such matching can facilitate proper attribution to content owners for example.
One type of matching is based on matching the melodies in a performance with the melodies in a previously recorded composition. Systems of such type are typically referred to as melody matching systems. In some cases, matching is particularly challenging due to differences in melodies in the compositions being analyzed. For example, in some cases, the melodies in the performance may be somewhat distinct from those in the previously-recorded composition. Distinctions can arise due to skill level of a user performing a composition, extensive improvisation by the user or for any number of other reasons. False positives occur when the system erroneously determines that the user performance and the previously-recorded composition are the same composition. False positives decrease the reliability of melody matching systems and, as such, are ideally minimized. Accordingly, systems and methods that reduce false positives, thereby enhancing the confidence of melody matching systems are desirable.